IgE actions on CD4+ T cells, mast cells, and macrophages participate in the pathogenesis of experimental abdominal aortic aneurysms

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Wang, J., J. S. Lindholt, G. K. Sukhova, M. A. Shi, M. Xia, H. Chen, M. Xiang, et al. 2014. “IgE actions on CD4+ T cells, mast cells, and macrophages participate in the pathogenesis of experimental abdominal aortic aneurysms.” EMBO Molecular Medicine 6 (7): 952-969. doi:10.15252/emmm.201303811. http:// dx.doi.org/10.15252/emmm.201303811.

Published Version doi:10.15252/emmm.201303811

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:12785965

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Research Article

IgE actions on CD4+ T cells, mast cells, and macrophages participate in the pathogenesis of experimental abdominal aortic aneurysms

Jing Wang1, Jes S Lindholt2, Galina K Sukhova1, Michael A Shi1, Mingcan Xia3, Han Chen1,4, Meixiang Xiang4, Aina He1, Yi Wang1, Na Xiong3, Peter Libby1, Jian-An Wang4,** & Guo-Ping Shi1,*

Abstract from vascular smooth muscle cells (SMCs) and endothelial cells (ECs) (Sun et al, 2007a,b); MC-specific proteases (e.g. chymase and (IgE) activates mast cells (MCs). It remains tryptase) to induce vascular cell apoptosis (Heikkila¨ et al, 2008; Sun unknown whether IgE also activates other inflammatory cells, and et al, 2009; Zhang et al, 2011; den Dekker et al, 2012); angiogenic contributes to the pathogenesis of abdominal aortic aneurysms factors (e.g. basic fibroblast growth factor) to stimulate angiogenesis (AAAs). This study demonstrates that CD4+ T cells express IgE (Lappalainen et al, 2004); and to induce EC inflammatory FceR1, at much higher levels than do CD8+ T cells. IgE responses (Dileepan & Stechschulte, 2006). Multiple mechanisms— induces CD4+ T-cell production of IL6 and IFN-c, but reduces their including lipoproteins, immunoglobulins, neurotrophic factors (e.g. À À production of IL10. FceR1 deficiency (Fcer1a / ) protects apolipo- substance P and neuron growth factor), stem cell factor, complements À À protein E-deficient (Apoe / ) mice from angiotensin-II infusion- (e.g. C3a and C5a), endotoxin, inflammatory (e.g. IL1 and induced AAAs and reduces plasma IL6 levels. Adoptive transfer of TNF-a) (Xu & Shi, 2012), or even acute restraint stress (Huang et al, CD4+ T cells (but not CD8+ T cells), MCs, and macrophages from 2002)—can mediate MC activation. Among these MC activators, À À À À À À Apoe / mice, but not those from Apoe / Fcer1a / mice, immunoglobulin E (IgE) is probably the best known (Ishizaka et al, À À À À increases AAA size and plasma IL6 in Apoe / Fcer1a / recipient 1978; Razin et al, 1983; Dvorak et al, 1985). Mast cells inactivation mice. Biweekly intravenous administration of an anti-IgE monoclo- by blocking IgE activity by use of anti-IgE has been inves- À À nal ablated plasma IgE and reduced AAAs in Apoe / tigated widely among patients with allergic asthma, food , mice. Patients with AAAs had significantly higher plasma IgE levels atopic dermatitis, atopic eczema, chronic autoimmune urticaria, and than those without AAAs. This study establishes an important role chronic rhinosinusitis (Milgrom et al, 1999; Leung et al, 2003; Busse of IgE in AAA pathogenesis by activating CD4+ T cells, MCs, and et al, 2011). We recently found that in addition to MCs, IgE also macrophages and supports consideration of neutralizing plasma targets macrophages, SMCs, and ECs, and induces and IgE in the therapeutics of human AAAs. chemokine production from these atherosclerosis-pertinent cells (Wang et al, 2011). This study demonstrates that T cells express IgE Keywords abdominal aortic aneurysm; IgE; macrophage; ; T cells high affinity receptor FceR1. Using angiotensin-II (Ang-II) infusion- À À Subject Categories Haematology; Immunology induced experimental AAAs in apolipoprotein E-deficient (Apoe / ) DOI 10.15252/emmm.201303811 | Received 31 December 2013 | Revised 21 mice, we tested whether IgE actions on T cells, MCs, and macrophag- May 2014 | Accepted 22 May 2014 | Published online 24 June 2014 es contribute to AAA pathogenesis and whether inhibition of IgE EMBO Mol Med (2014) 6: 952–969 activity using FceR1-deficient mice or ablation of plasma IgE using anti-IgE monoclonal antibody (mAb) ameliorates AAA development. Introduction Results Mast cells (MCs) contribute importantly to the pathogenesis of atherosclerosis and abdominal aortic aneurysms (AAAs) (Kovanen, FceR1 expression on CD4+ and CD8+ T cells 2007; Sun et al, 2007a,b; Bot & Biessen, 2011; Swedenborg et al, 2011). After activation, MCs release pro-inflammatory cytokines Inflammatory cells express the high-affinity IgE receptor FceR1, e.g. (e.g. IFN-c and IL6) to induce matrix-degrading protease expression MCs, , eosinophils, dendritic cells, Langerhans cells,

1 Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA 2 Department of Cardiovascular and Thoracic Surgery, Elitary Research Centre of Individualized Medicine in Arterial Diseases, University Hospital of Odense, Odense, Denmark 3 Department of Veterinary and Biomedical Sciences, The Pennsylvania State University, University Park, PA, USA 4 Cardiovascular Key Lab of Zhejiang Province, Department of Cardiology, College of Medicine, The Second Affiliated Hospital, Zhejiang University, Hangzhou, China *Corresponding author. Tel: +1 617 525 4358; Fax: +1 617 525 4380; E-mail: [email protected] **Corresponding author. Tel: +86 138 057 86328; Fax: +86 571 87022776; E-mail: [email protected]

952 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors. Published under the terms of the CC BY 4.0 license Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

macrophages, and (Baniyash et al, 1986; Shibaki et al, (Fig 1E). Cell culture medium ELISA showed that IgE activation 1996; Boesiger et al, 1998; Dombrowicz et al, 2000; Katoh et al, dose-dependently reduced CD4+ T-cell IL10 production (Fig 1F). 2000; Grayson et al, 2007; Mancardi et al, 2008). Neurons and Consistent with the differences in FceR1 expression levels between vascular SMCs and ECs also express FceR1 (Andoh & Kuraishi, CD4+ and CD8+ T cells, cellular IFN-c mRNA levels and medium IL6 2004; Wang et al, 2011), suggesting a broad spectrum of target cells protein levels were significantly higher in CD4+ T cells than in CD8+ for IgE. IgE activates MCs by targeting FceR1 (Xu & Shi, 2012). We T cells at various IgE concentrations. In contrast, we detected minimal have recently shown that IgE also activates macrophages via inter- levels of mRNA and medium IL10 in CD8+ T cells (Fig 1E and F). action with cell surface FceR1 (Wang et al, 2011). This study used Why IgE promotes MC and survival and proliferation CD4+ and CD8+ T cells selected from splenocytes from wild-type (Kawakami & Galli, 2002), but induces macrophage apoptosis (Wang C57BL/6 mice by nylon wool column purification and consequently et al, 2011), remains unknown. Altered cytokine expression from anti-I-Ab antibody depletion (Kokkinopoulos et al, 1992), followed IgE-treated CD4+ and CD8+ T cells may result from IgE actions affect- by anti-CD4 and anti-CD8 antibody-coated magnetic column positive ing T-cell survival or proliferation. To test this possibility, we cultured selection. Cell purity ranged from 94 to 98% (Supplementary Fig both CD4+ and CD8+ T cells in anti-CD3 (1 lg/ml) and anti-CD28 S1). To examine whether T cells also express FceR1, we treated both (1 lg/ml) monoclonal antibodies and stimulated these cells with IgE CD4+ and CD8+ T cells with IgE, inflammatory cytokines (IFN-c, (50 lg/ml) for 3 days. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra- IL6, and TNF-a), or Ang-II. Real-time polymerase chain reaction zolium bromide (MTT) cell survival assay and medium IL2 ELISA (RT-PCR) demonstrated the expression of mRNAs that encode all demonstrated that IgE did not affect either survival or proliferation of À À three FceR1 subunits—FceR1a,FceR1b, and FceR1c—in CD4+ T CD4+ or CD8+ T cells from either Apoe / mice or FceR1a-deficient À À À À cells, but at much lower levels in CD8+ T cells. Cell stimulation Apoe / Fcer1a / mice (Supplementary Fig S4A–D). with IgE, cytokines, or Ang-II did not significantly change FceR1 À À expression from either CD4+ or CD8+ T cells (Fig 1A). To deter- Absence of FceR1 reduces experimental AAAs in Apoe / mice mine what percentages of these T cells express FceR1, we gated total CD3+ T cells from mouse splenocytes and selected CD4+ and CD8+ Chronic subcutaneous infusion of Ang-II induces experimental À À T cells to detect the percentage of FceR1+ cells in these two popula- AAAs in Apoe / mice (Daugherty et al, 2000; Schulte et al, 2010; tions of T cells (Supplementary Fig S2). Flow cytometry (FACS) Zhang et al, 2012). We first tested whether AAA development in À À analysis of splenocytes using anti-FceR1a-biotin mAb demonstrated Apoe / mice associates with plasma IgE level changes. Among 16 À À comparable percentages of CD4+ and CD8+ T cells that express Apoe / mice that received Ang-II infusion, 13 developed AAAs and surface FceR1a, but negligible intracellular FceR1a after subtracting 3 did not. Mice with confirmed AAAs had significantly higher the baseline signals from the antibody controls (Fig 1B). plasma IgE levels than those before Ang-II infusion or those that did Supplementary Fig S3 shows a representative FACS analysis. To not develop AAA even after receiving Ang-II (Fig 2A), although compare the relative expression of the same three FceR1 subunits in their blood pressures or heart rates did not differ significantly CD4+ T cells to those of cells known to express FceR1, we (Supplementary Fig S5). Increased plasma IgE may trigger inflam- performed RT-PCR from both bone marrow-derived MC (BMMCs) matory cytokine expression from T cells, MCs, macrophages, and and macrophages and found that BMMCs and macrophages other inflammatory cells, thereby promoting AAA pathogenesis. expressed much higher levels of the FceR1a subunit, but compara- Interruption of IgE activity may thus attenuate AAA growth. To test ble levels of FceR1b (BMMCs: 0.08–0.17%; macrophages: <0.001– this hypothesis, we produced Ang-II infusion-induced AAAs in both À À À À À À 0.025% relative to b-) and FceR1c (BMMCs: 0.40–0.80%; FceR1-sufficient Apoe / mice and FceR1-deficient Apoe / Fcer1a / À À macrophages: 0.09–1.27% relative to b-actin) subunits to those of mice. In these experiments, Apoe / mice tended to hvae a higher À À À À CD4+ T cells, depending on the stimulus (Fig 1C). Immunoblot post-Ang-II infusion mortality rate than that of Apoe / Fcer1a / analysis showed that FceR1 protein expression was also much mice (40% versus 20%, P = 0.314, Fig 2B). To achieve sufficient higher in CD4+ T cells than in CD8+ T cells and that IgE, cytokines, numbers of mice that survived from Ang-II infusion and for AAA À À À À À À and Ang-II did not affect FceR1a expression (Fig 1D), concordant lesion analysis, we used 30 Apoe / mice and 15 Apoe / Fcer1a / À À À À À À the results of RT-PCR and FACS analysis (Fig 1A and B). These data mice to provide 18 Apoe / and 12 Apoe / Fcer1a / survivors À À À À suggest that CD4+ T cells expressed much higher levels of FceR1 after 28 days of Ang-II infusion. Apoe / Fcer1a / mice showed a À À (RT-PCR) than did CD8+ T cells, although similar percentages of much lower AAA incidence than did Apoe / mice (33.3% versus these T cells expressed FceR1 (FACS). We have previously shown 83.3%, P = 0.008, Fig 2B). Suprarenal maximal outer aortic diame- that either aggregated (SPE-7) or monomeric (H1 DNP-e-206) forms ters measured in situ from anesthetized mice were significantly À À À À À À of IgE have equal potency as macrophage activators, and that IgE smaller in Apoe / Fcer1a / mice than in Apoe / mice (Fig 2B). À À À À À À did not affect this activity of IgE (Wang et al, 2011). To test As we anticipated, Apoe / and Apoe / Fcer1a / mice that devel- whether FceR1 on CD4+ and CD8+ T cells was functional and also oped AAAs had significantly higher plasma IgE levels than did those independent of antibody cross-linking, we demonstrated that SPE-7 did not develop AAAs, although FceR1-deficiency enhanced plasma IgE produced a concentration-dependent increase in the expression IgE levels post-Ang-II infusion, independent of AAA formation. In of IL6 and IFN-c mRNA and protein production by CD4+ and CD8+ contrast, plasma IgG or IgM levels did not change significantly in T cells, as determined by RT-PCR (Fig 1E) and by culture medium mice with and without AAAs from either genotype (Fig 2C). AAA ELISA (Fig 1F). In contrast, CD4+ T cells showed biphasic but insig- lesion (macrophage-positive area, T-cell content, nificant changes in IL10 mRNA levels after exposure to low concen- content, major complex [MHC] trations of IgE (0–10 lg/ml), but demonstrated significant reduction class-II content, and chemokine chemoattractant protein- of IL10 expression under high IgE concentrations (50–100 lg/ml) 1 [MCP-1] content), lesion cell proliferation (Ki67-positive areas),

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 953 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Figure 1.

954 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

À À À À microvessel number (CD31), apoptotic cell number, media elastin pient mice much greater than those from Apoe / Fcer1a / mice, fragmentation, and presence of SMC in the tunica media (loss of even though they had similar levels of plasma IgE (Fig 3C). Suprare- À À media SMCs) were all significantly lower in AAA lesions from nal aortic diameters were also significantly reversed in Apoe / À À À À À À À À À À Apoe / Fcer1a / mice than in lesions from Apoe / mice (Fig 2D Fcer1a / mice after receiving CD4+ T cells from Apoe / mice, but À À À À and E). However, FceR1 deficiency did not significantly affect body not those from Apoe / Fcer1a / mice (Fig 3C). Reconstitution of À À À À weight, diastolic and systolic blood pressures, heart rate, plasma CD4+ T cells from Apoe / Fcer1a / mice did not affect recipient lipid profiles (Supplementary Fig S6A–E) before or after AAA mice lesion macrophage contents (Fig 3D, left panel), but lesion total production, or AAA lesion total interstitial collagen contents CD4+ T-cell numbers did not differ significantly from those receiving À À (Supplementary Fig S6F). These observations establish an important CD4+ T cells from Apoe / mice (Fig 3D, middle panel)—suggesting role of IgE in AAA pathogenesis. that absence of FceR1 on donor CD4+ T cells did not affect AAA lesion T-cell infiltration, but reduced macrophage accumulation. IgE actions on CD4+ T cells are essential to AAAs Therefore, the FceR1 on CD4+ T cells may be permissive for macro- phage infiltration into AAA lesions, similar to the prior conclusion Increased plasma IgE in AAA mice (Fig 2A and C), reduced AAAs in that M1 macrophage infiltration in white adipose tissues requires À À À À FceR1-deficient Apoe / Fcer1a / mice (Fig 2B), and increased IL6 prior CD8+ T-cell activation (Nishimura et al, 2009). Reduced and IFN-c and reduced IL10 in CD4+ T cells after IgE activation macrophages in AAA lesions from recipient mice receiving CD4+ T À À À À (Fig 1E and F) suggest a role of IgE actions on CD4+ T cells in AAAs. cells from Apoe / Fcer1a / mice yielded lower levels of MHC class- To test this hypothesis, we adoptively transferred equal numbers II—marker of inflammation (Zhang et al, 2012)—in these AAA À À (1 × 107 per recipient) of splenic CD4+ T cells from age-matched lesions, compared with those receiving CD4+ T cells from Apoe / À À À À À À À À À À Apoe / mice and Apoe / Fcer1a / mice to Apoe / Fcer1a / reci- mice, although such difference did not reach statistical significance pient mice, followed by Ang-II infusion to elicit AAA formation. After (Fig 3D, right panel). À À À À 28 days, we confirmed the presence of FceR1a-positive donor cells in In contrast, recipient Apoe / Fcer1a / mice receiving CD8+ T À À À À À À À À À À AAA lesions of recipient Apoe / Fcer1a / mice receiving CD4+ T cells from Apoe / mice or Apoe / Fcer1a / mice did not change À À À À À À cells from Apoe / mice (Fig 3A, left two panels) using a hamster AAA formation in Apoe / Fcer1a / recipient mice. Anti-FceR1a À À À À anti-mouse FceR1a antibody. Donor cells from Apoe / Fcer1a / mAb immunostaining and immunofluorescent co-staining together mice were used as negative controls for immunostaining (Fig 3A, with an anti-Ki67 rabbit polyclonal antibody of AAA lesion sections À À À À right panel). Antibody specificity was further tested using AAA lesion from Apoe / Fcer1a / recipient mice receiving donor CD8+ T cells À À consecutive sections with or without the primary antibody and using from Apoe / mice showed the presence of donor CD8+ T cells in À À À À AAA sections from the Apoe / Fcer1a / mice (Supplementary Fig AAA lesions (Supplementary Fig S8B and S9A, left two panels) and S7). Low magnifications (20× or 40×) showing the entire aortas of these donor cells also underwent clonal expansion (Supplementary À À À À À À Apoe / Fcer1a / recipient mice receiving donor cells from Apoe / Fig S9B). Sections from recipient mice receiving CD8+ T cells from À À À À À À À À mice and Apoe / Fcer1a / mice are illustrated in Supplementary Fig Apoe / Fcer1a / mice served as negative controls for the FceR1a À À S8A. Immunofluorescent co-staining of AAA lesions from Apoe / mAb (Supplementary Fig S8B and S9A, right panel). Unlike donor À À À À À À Fcer1a / recipient mice receiving donor CD4+ T cells from Apoe / CD4+ T cells, donor CD8+ T cells from either Apoe / mice À À À À mice for FceR1a and Ki67 demonstrated that donor CD4+ T cells not or Apoe / Fcer1a / mice did not increase plasma IgE levels only targeted to the AAA lesions, but also underwent clonal expan- significantly (Supplementary Fig S9C). Although we detected mild À À sion (Fig 3B). As in Cd3e / mice (Milner et al, 2007), CD4+ T-cell increases in AAA incidence in all mice that received donor CD8+ À À À À À À reconstitution increased plasma IgE levels in recipient mice, but we T cells from Apoe / mice or Apoe / Fcer1a / mice, and there did not detect any differences between the two donor cells from were complete reverses of the post-Ang-II mortality rate in mice À À À À À À À À Apoe / mice and Apoe / Fcer1a / mice (Fig 3C). Therefore, that received CD8+ T cells from Apoe / mice but not from À À À À À À CD4+ T-cell-induced plasma IgE level increase did not depend on Apoe / Fcer1a / mice, suprarenal aortic diameters from Apoe / À À À À FceR1 expression, although the mechanism by which CD4+ T cells Fcer1a / mice that received donor cells from Apoe / mice or À À À À augment recipient mice IgE levels remains incompletely understood Apoe / Fcer1a / mice were not different from those of the À À À À (Milner et al, 2007). Although not statistically significant, CD4+ T parental Apoe / Fcer1a / mice (Supplementary Fig S9C). À À cells from Apoe / mice partially or fully restored both the AAA inci- Consistent with unchanged aortic diameters, AAA lesion macro- À À À À À À À À dence rate and post-Ang-II mortality rate in Apoe / Fcer1a / reci- phage content did not differ between Apoe / Fcer1a / mice and

Figure 1.FceR1 expression and activity in cells treated with or without different stimuli. ◂ e 1 e 1a e 1b e 1c b 4+ 8+ A RT-PCR to determine mRNA levels of three Fc R subunits, Fc R ,Fc R , and Fc R relative to -actin from CD and CD T cells. B FACS analysis to determine the percentage of FceR1a-expressing T cells by staining surface and intracellular FceR1a on CD4+ and CD8+ T cells after subtracting the background signals from the IgG isotypes. C RT-PCR to determine mRNA levels of three FceR1 subunits relative to b-actin from bone marrow-derived mast cells (BMMCs) and macrophages. D Immunoblot analysis of FceR1a on CD4+ and CD8+ T cells. Both CD4+ and CD8+ T-cell immunoblots were from the same SDS–PAGE. E RT-PCR to determine cytokine (IL6, IFN-c, and IL10) mRNA level changes in CD4+ and CD8+ T cells, treated with or without different doses of IgE (SPE-7) as indicated. P values refer to comparisons to untreated cells (IgE = 0 lg/ml). F ELISA to determine culture medium cytokine (IL6, IFN-c, and IL10) levels in CD4+ and CD8+ T cells, treated with and without different doses of IgE as indicated. P values in black refer to comparisons to untreated cells (IgE = 0 lg/ml). P values in red refer to comparisons between CD4+ and CD8+ T cells. Data information: Data in panels A-E are mean Æ SEM from 3–5 independent experiments.

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 955 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Figure 2. Role of FceR1 in Ang-II infusion-induced AAAs. A ELISA determined plasma IgE levels in ApoeÀ/À mice before and 28 days after Ang-II infusion with and without AAA formation. B Post-Ang-II infusion rupture-associated mortality rate calculated from survived mice versus total mice used (indicated), and AAA incidence and maximal suprarenal outer aortic diameter from in situ calculated from all survived mice. C Plasma IgE, IgG, and IgM levels in survived mice with and without AAAs. D, E AAA lesion macrophage content, CD4+ T-cell content, dendritic cell content, major histocompatibility complex (MHC) class-II-positive area, and chemokine MCP-1- positive area (D), and AAA lesion Ki67-positive proliferating cell number, CD31-positive microvessel number, TUNEL-positive apoptotic cell area, arterial wall elastin fragmentation grade, and media smooth muscle cell (SMC) accumulation grade (E) from both ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice harvested at 28 days after Ang-II infusion. Data information: Data are mean Æ SEM. Number of mice per group is indicated in each bar.

À À those receiving CD8+ T-cell adoptive transfer (Supplementary Fig receiving donor cells from Apoe / mice, but did not reach statisti- À À S9D, left panel), although lesion T-cell content and MHC class-II- cal significance from those receiving donor cells from Apoe / À À positive areas increased in AAA lesions from recipient mice Fcer1a / mice (Supplementary Fig S9D, right two panels). Our

956 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

Figure 3. IgE actions on CD4+ T cells in AAAs. A Anti-FceR1a antibody-mediated immunostaining to detect donor CD4+ T cells from ApoeÀ/À mice (left two panels) and Fcer1aÀ/ÀApoeÀ/À mice (right panel) in recipient Fcer1aÀ/ÀApoeÀ/À mice. Scale: 200 lm; insert scale: 50 lm. BFceR1a and Ki67 antibody-mediated immunofluorescent co-staining to detect proliferating donor CD4+ T cells from ApoeÀ/À mice in recipient Fcer1aÀ/ÀApoeÀ/À mice. Scale: 200 lm; insert scale: 60 lm. C, D (C) Plasma IgE level, AAA incidence, post-Ang-II infusion aortic rupture-associated mortality rate, and suprarenal maximal outer aortic diameter; and (D) AAA lesion macrophage content, CD4+ T-cell content, and major histocompatibility complex (MHC) class-II-positive area from ApoeÀ/À mice, Fcer1aÀ/ÀApoeÀ/À mice, and Fcer1aÀ/ÀApoeÀ/À recipient mice receiving donor CD4+ T cells from ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice. Data are mean Æ SEM. Number of mice per group is indicated in each bar.

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 957 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

À À À À À À À À observations from Apoe / Fcer1a / mice reconstituted with CD4+ receiving macrophages from Apoe / Fcer1a / mice served as and CD8+ T cells suggest that IgE effects on CD4+ T cells influenced negative controls for immunostaining (Supplementary Fig S8D and AAA pathogenesis to a much greater extent than did CD8+ Tcells. Fig 5A, right panel). Adoptive transfer of macrophages from either À À À À À À Apoe / or Apoe / Fcer1a / mice did not significantly change AAA formation requires IgE actions on MCs and macrophages plasma IgE levels (Fig 5B), but IgE actions on macrophages contrib- À À uted to AAA formation. Donor macrophages from Apoe / mice, À À À À The foregoing experiments established an important role of IgE on but not those from Apoe / Fcer1a / mice, increased recipient AAA CD4+ T cells, but both MCs and macrophages also express FceR1 incidence and post-Ang-II mortality rate. Donor macrophages from À À À À À À (Fig 1C) (Boesiger et al, 1998; Wang et al, 2011). IgE activates MCs Apoe / mice, but not those from Apoe / Fcer1a / mice, also and induces MC and release of granule-associated reversed outer aortic diameters in recipient mice, although such change inflammatory mediators (Ishizaka et al, 1978; Razin et al, 1983; did not reach statistical significance (Fig 5B). Reconstitution of macro- À À À À À À Dvorak et al, 1985). IgE also activates macrophages to release cyto- phages from Apoe / mice, but not those from Apoe / Fcer1a / kines and chemokines or to undergo apoptosis (Wang et al, 2011). mice, also increased lesion macrophage-positive area, T-cell IgE actions on these inflammatory cells therefore may also contrib- number, and MHC class-II-positive area (Fig 5C), establishing a role ute to AAA pathogenesis as do those from CD4+ T cells. of IgE actions on macrophages in AAA lesions. To examine the role of IgE activity on MCs in AAAs, we prepared À À À À À À BMMCs from both Apoe / and Apoe / Fcer1a / mice and adop- Anti-IgE monoclonal antibody administration limits experimental À À À À tively transferred them intravenously into Apoe / Fcer1a / reci- AAA formation pient mice, followed by Ang-II infusion to elicit AAA formation. Anti-FceR1a mAb immunostaining demonstrated the anticipated Observations from this study suggest that IgE actions on CD4+ T À À accumulation of donor FceR1a-positive BMMCs from Apoe / mice cells (Fig 3), MCs (Fig 4), and macrophages (Fig 5) contribute inde- À À À À in Apoe / Fcer1a / recipient mouse AAA lesions (Supplementary pendently to experimental AAAs to different extents. These observa- Fig S8C and Fig 4A, left two panels). The same immunostaining of tions agree with prior studies that CD4+ T cells, MCs, and AAA sections from recipient mice that received BMMCs from macrophages or their activation contribute directly and indepen- À À À À Apoe / Fcer1a / mice tested antibody specificity (Supplementary dently to AAA formation (Xiong et al, 2004, 2009; Sun et al, 2007b). Fig S8C and Fig 4A, right panel). Similar to results obtained with Interference with IgE signaling by genetically depleting the high- À À À À CD4+ T-cell-reconstituted Apoe / Fcer1a / recipient mice, BMMC affinity IgE receptor FceR1 (Fig 2), or IgE neutralization using anti- reconstitution increased recipient mice plasma IgE levels, but did not IgE antibodies, therefore could suppress AAA growth. To test this À À differ between the two donor cells (Fig 4B). BMMCs from Apoe / hypothesis, we administered an anti-IgE mAb intravenously 1 day À À À À mice, but not those from Apoe / Fcer1a / mice, increased AAA inci- before initiating the AAA (Ang-II mini-pump implantation), and dence, post-Ang-II mortality rate, and maximal outer aortic diameter once more at 2 weeks after AAA initiation using doses previously À À À À in Apoe / Fcer1a / recipient mice (Fig 4B). Statistically insignificant established as effective in allergic mice and humans (Coyle et al, differences in AAA incidences and post-Ang-II mortality might be due 1996; Haile et al, 1999; Milgrom et al, 1999; Leung et al, 2003; to our relatively small sample sizes. AAA lesion apoptotic cell number Busse et al, 2011). Two doses of anti-IgE antibody did not change (Fig 4C) and microvessel number (Fig 4D) were also restored partially body weight before or after AAA formation (Fig 6A), but signifi- À À in recipient mice receiving BMMCs from Apoe / mice, but not receiv- cantly reduced plasma IgE by more than 70% (Fig 6B). IgE ablation À À À À ing those from Apoe / Fcer1a / mice. Surprisingly, reconstitution of with anti-IgE antibody reduced AAA incidence, post-Ang-II mortality À À À À À À MCs—whether from Apoe / mice or Apoe / Fcer1a / mice—did rate, and outer aortic diameters (Fig 6B), although the differences in not significantly increase recipient mice AAA lesion macrophage AAA incidence and poat-Ang-II mortality rate did not reach statisti- content, while lesion T cells, MHC class-II-positive areas, or lesion cell cal significance (Fisher’s exact test). proliferation all tended to recover after BMMCs adoptive transfers, Our leukocyte (CD4+ T cells, MCs, or macrophages) reconstitu- although these trends did not achieve statistical significance (Supple- tion experiments showed that IgE actions on each of these cell types mentary Fig S10). These data suggest that IgE actions on MCs have contributed directly, but to different extents, to AAA formation. As limited impact on macrophage and T-cell infiltration in AAA lesions. we anticipated, anti-IgE antibody treatment, which should block IgE Macrophages in AAA lesions come from blood-borne monocytes, activities on all these target cells, as well as intrinsic vascular cells but a short lifespan after repopulation in recipient mice (Leuschner (Wang et al, 2011), significantly suppressed AAA lesion macro- et al, 2012) made it technically difficult to study donor monocytes phage content, CD4+ T-cell number, MHC class-II-positive area, in AAA recipient mice. Instead, we used macrophages, which do lesion cell apoptosis, microvessel number, and lesion cell prolifera- successfully repopulate in recipient mice to study macrophage acti- tion (Fig 6C–F). Non-selective targeting of IgE actions using anti-IgE vation in experimental AAA (Xiong et al, 2009). FACS analysis antibody therapy thus yielded similar effects in suppressing AAA confirmed that peritoneal macrophages prepared from donor mice pathogenesis (Fig 6) to those of genetic depletion of IgE receptor contained > 95% macrophages after an adhesion selection to deplete FceR1 (Fig 2). other cells (Supplementary Fig S11). Donor macrophages in AAA À À À À lesions from recipient Apoe / Fcer1a / mice were localized by IgE actions alter levels of inflammatory cytokines in plasma immunostaining frozen AAA sections with anti-FceR1a mAb. À À FceR1a-positive donor macrophages from Apoe / mice appeared in One of the most important activities of IgE on inflammatory cells is À À À À recipient Apoe / Fcer1a / mouse AAA lesions (Supplementary Fig to induce their pro-inflammatory cytokine production, which may S8D and Fig 5A, left two panels). Lesions from recipient mice accelerate AAA formation. Reduced AAAs after interruption of IgE

958 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

Figure 4. IgE actions on bone marrow-derived mast cells (BMMCs) in AAAs. A Anti-FceR1a antibody-mediated immunostaining to detect donor BMMCs from ApoeÀ/À mice (left two panels) and Fcer1aÀ/ÀApoeÀ/À mice (right panel) in recipient Fcer1aÀ/ÀApoeÀ/À mice. Scale: 200 lm, insert scale: 50 lm. B Plasma IgE level, AAA incidence, post-Ang-II infusion aortic rupture-associated mortality rate, and suprarenal maximal outer aortic diameter. C, D (C) AAA lesion TUNEL-positive apoptotic cell percentage; and (D) AAA lesion CD31-positive microvessel number ApoeÀ/À mice, Fcer1aÀ/ÀApoeÀ/À mice, and Fcer1aÀ/À ApoeÀ/À recipient mice receiving donor CD4+ T cells from ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice. Representative AAA lesions from adoptive transferred mice for panels (C) and (D) are shown to the right. Scale: 200 lm. Data information: (B-D) Data are mean Æ SEM. Number of mice per group is indicated in each bar. interactions with inflammatory cells using either FceR1-deficient indexed aortic diameters (r = 0.285, P = 0.002, linear regression) mice or cells, or anti-IgE antibodies, may associate with reduced (Rohde et al, 1999). IL6 deficiency or antibody depletion of plasma circulating cytokine levels. In AAA patients, plasma IL6 levels range IL6 protects mice from aortic elastase perfusion-induced AAAs in from 20–40 pg/ml and correlate positively and significantly with mice (Thompson et al, 2006), suggesting an important role of IL6 in

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 959 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Figure 5. IgE actions on macrophages in AAAs. A Anti-FceR1a antibody-mediated immunostaining to detect donor macrophages from ApoeÀ/À mice (left two panels) and Fcer1aÀ/ÀApoeÀ/À mice (right panel) in recipient Fcer1aÀ/ÀApoeÀ/À mice. Scale: 200 lm, insert scale: 50 lm. B, C (B) Plasma IgE level, AAA incidence, post-Ang-II infusion aortic rupture-associated mortality rate, and suprarenal maximal outer aortic diameter; and (C). AAA lesion macrophage content, CD4+ T-cell content, and major histocompatibility complex (MHC) class-II-positive area from ApoeÀ/À mice, Fcer1aÀ/ÀApoeÀ/À mice, and Fcer1aÀ/ÀApoeÀ/À recipient mice receiving donor CD4+ T cells from ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice. Data are mean Æ SEM. The number of mice per group is indicated in each bar.

À À À À À À AAAs. This study demonstrated that plasma IL6 levels in Apoe / reduction in plasma IL6 levels, to the levels from Apoe / Fcer1a / mice resemble those in AAA patients, and fell significantly in mice (Fig 7F). These observations indicate the prominence of IL6 as a À À À À Apoe / Fcer1a / mice (Fig 7A). The sensitivity of the ELISA used mediator released after IgE activation of inflammatory cells. In contrast, for IFN-c did not permit detection and analysis of plasma levels of although IgE suppressed IL10 production from isolated CD4+ Tcellsin this important pro-inflammatory cytokine also implicated in AAA vitro (Fig 1E and F), we did not detect significant differences in plasma À À À À À À pathogenesis (King et al, 2009). Low levels of plasma IFN-c in these IL10 levels between Apoe / and Apoe / Fcer1a / mice, or those mice agreed with the observations that T cells produced 10-fold less that received different types of CD4+ T cells, CD8+ T cells, MCs, or IFN-c than IL6 after IgE stimulation (Fig 1F). Reconstitution of CD4+ macrophages (Supplementary Fig S12A–E). Nor did anti-IgE anti- T cells and MCs, but not CD8+ T cells and macrophages, from body affect plasma IL10 levels in mice with experimental AAAs À À Apoe / mice significantly restored plasma IL6 levels (Fig 7B–E). (Supplementary Fig S12F), although IgE neutralization may affect Plasma IL6 levels in recipient mice receiving macrophages from other untested circulating anti-inflammatory cytokines. À À À À Apoe / Fcer1a / mice were significantly lower than those from À À recipient mice receiving macrophages from Apoe / mice (Fig 7E). Increased plasma IgE levels in patients with AAAs À À While Apoe / mice receiving IgG treatment showed comparable À À plasma IL6 levels to those of untreated Apoe / mice, those receiv- The use of IgE receptor FceR1-deficient mice, adoptive transfer of ing two doses of an anti-IgE antibody demonstrated significant various inflammatory cell types, and anti-IgE antibody therapy

960 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

Figure 6. Anti-IgE antibody treatment of experimental AAAs. A Body weight in ApoeÀ/À mice that received control IgG and anti-IgE mAb before and after AAA production. B Plasma IgE level, AAA incidence, post-Ang-II infusion aortic rupture-associated mortality rate, and suprarenal maximal outer aortic diameter. C–F (C) Lesion macrophage content, CD4+ T-cell content, and major histocompatibility complex (MHC) class-II-positive area; (D) Lesion TUNEL-positive apoptotic cell content; (E) CD31-positive microvessel number; and (F) Lesion Ki67-positive proliferating cell number from ApoeÀ/À mice receiving control IgG or anti-IgE mAb. Data information: Data are mean Æ SEM. The number of mice per group is indicated in each bar. Representative figures for panels C to E are shown to the right. Scale: 50 lm. established that IgE aggravates experimental AAA formation by acti- have previously reported that AAA patients have significantly higher vating mononuclear leukocytes. Mice have increased plasma IgE plasma MC chymase and tryptase levels than controls (Sun et al, under conditions that produce AAAs (Fig 2A and C, left panel). We 2009; Zhang et al, 2011). We therefore measured plasma IgE levels

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 961 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Figure 7. Plasma IL6 levels from different groups of mice as determined by ELISA. A–F Plasma IL6 levels in ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice (A) and Fcer1aÀ/ÀApoeÀ/À mice that received adoptive transfer of CD4+ T cells (B), CD8+ T cells (C), bone marrow-derived mast cells (BMMCs) (D), and macrophages (E) from ApoeÀ/À and Fcer1aÀ/ÀApoeÀ/À mice, and ApoeÀ/À mice that received biweekly intravenous administration of anti-IgE mAb or corresponding IgG control (F). Data are mean Æ SEM. The number of mice per group is indicated in each bar.

from 487 male AAA patients and compared them by ELISA with between AAA patients and controls (AUC [area under the ROC those from 200 sex-matched and age-matched individuals without curve] = 0.588, P < 0.001), with optimal sensitivity of 0.60 and AAAs. Plasma IgE levels showed skewed distribution, therefore data specificity of 0.59 (Supplementary Fig S13). Among this population, analysis used the non-parametric Mann–Whitney test. AAA patients patients with peripheral arterial disease (PAD) also had significantly had significantly higher plasma IgE levels than controls higher plasma IgE levels than controls (250.79 Æ 229.88 ng/ml (81.79 Æ 61.65 ng/ml versus 7.10 Æ 1.70 ng/ml, mean Æ SD, versus 15.32 Æ 3.33 ng/ml, mean Æ SD, P < 0.001) (Supplemen- P < 0.001) (Supplementary Table S1), although we were unable to tary Table S1). ascertain whether there were any differences in the events of type I , such as allergic asthma, allergic conjunctivitis, allergic Increased IgE and FceR1 expression in human AAA lesions rhinitis, , angioedema, urticaria, , penicillin allergy, cephalosporin allergy, and between the two We have previously shown that IgE activates human macrophages, populations because most of these clinical symptoms were diag- SMCs, and ECs, and induces their apoptosis (Wang et al, 2011). The nosed in community general practices. Receiver–operator character- present study suggested a role of IgE stimulation of CD4+ T cells, istic (ROC) curve analyses demonstrated that IgE levels discriminate MCs, and macrophages in experimental AAA formation. Inflammatory

962 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

cells in human AAA lesions might also express IgE receptor FceR1 subtype in experimental AAAs. The current study demonstrated and bind for IgE. Indeed, CD68-positive macrophages demonstrated FceR1 expression on CD4+ and CD8+ T cells and a direct role of IgE expression of FceR1 and IgE in consecutive frozen sections of on T-cell activation. CD4+ T cells expressed much more FceR1 mRNA human AAAs. Consistent with our prior study (Wang et al, 2011), and proteins than did CD8+ T cells (Fig 1A, B and D), leading to IgE-positive and FceR1-positive macrophages in human AAA lesions differences in IL6 and IFN-c expression in response to IgE stimulation also underwent apoptosis as shown by TUNEL-positivity (Fig 8A). (Fig 1E and F). Our data indicate a role for IL6 from CD4+ Tcellsin À À À À À À IgE binding to macrophages also triggers cytokine expression from AAAs. Plasma IL6 levels in Apoe / mice, Apoe / Fcer1a / mice, À À these cells (Wang et al, 2011) as in CD4+ T cells. Immunofluores- and those that received CD4+ and CD8+ TcellsfromApoe / mice À À À À cent staining colocalized the expression of both IgE and FceR1 on and Apoe / Fcer1a / mice (Fig 7B and C), changed concordantly CD4+ T cells in human AAA lesions (Fig 8B). with AAA lesion sizes (Fig 3C and Supplementary Fig S9C). In vitro IgE may also induce SMC apoptosis (Wang et al, 2011) in AAA experiments showed that IgE suppresses CD4+ T-cell IL10 expression lesions. While regions of human AAA specimens rich in a-actin- (Fig 1E and F), but we did not document significant change in plasma À À À À À À positive SMCs lacked detectable IgE signal (Fig 8C), many of the IL10 levels between Apoe / and Apoe / Fcer1a / mice, or those sparse SMCs in other regions bore IgE, observation consistent with that received different types of CD4+ or CD8+ Tcells(Supplementary a role for IgE in SMC depletion by apoptosis in human AAAs Fig S12A–C). Nonetheless, local level of IL10 in AAA lesions influ- (Fig 8D). Therefore, FceR1-positive SMCs in AAA lesions may be enced by IgE could occur. The observation of inverse production of particularly susceptible to apoptosis (Fig 8E). Some ECs in human IL6 and IL10 in human AAA lesion explant cultures, but not in AAA lesions also display immunoreactive FceR1 and IgE (Fig 8F), plasma from AAA patients (Vucevic et al, 2012), supports this suggesting that IgE signaling might also contribute to EC death hypothesis. (Wang et al, 2011) and intimal erosion, a possibility that merits IgE activation of MCs leads to MC degranulation and release of further investigation. inflammatory mediators (Gilfillan & Tkaczyk, 2006; Theoharides et al, 2007), including cytokines, chemokines, and proteases. We have previously demonstrated important roles of MCs and MC- Discussion derived IL6, IFN-c, MCP-1, chymases and tryptases in three indepen- dent experimental AAA preservations. By releasing these mediators, This study establishes a role for IgE in experimental AAAs. Using MCs promote angiogenesis, monocyte migration, and vascular SMC À À Ang-II infusion-induced experimental AAAs in Apoe / mice, apoptosis (Sun et al, 2007b, 2009; Zhang et al, 2011, 2012). IgE activa- FceR1-deficient mice, adoptive transfer of leukocyte populations, tion of MCs may enhance the production of these MC mediators and À À À À and anti-IgE mAb administration, we demonstrated that IgE actions consequent AAA pathogenesis. Apoe / Fcer1a / mice receiving MCs À À À À À À on CD4+ T cells, MCs, and macrophages, and possibly other inflam- from Apoe / mice but not those from Apoe / Fcer1a / mice have matory and vascular cells contribute to AAA pathogenesis. This increased AAA lesion apoptosis and microvascularization (Fig 4C and study implicates release of pro-inflammatory cytokines after IgE D), consistent with a regulatory role of IgE on MCs. As an indirect activation of leukocytes a major mechanism for affecting AAAs mechanism, MCs release IL6 and IFN-c after activation to induce among other possible pathways. vascular SMC and EC expression of cysteinyl cathepsins (Sun et al, The present observations show a direct role of IgE in T cells. T- 2007b), which also participate in arterial wall remodeling and AAA cell functions in AAAs have been controversial, although, it is diffi- pathogenesis (Sun et al, 2011, 2012; Qin et al, 2012). Indeed, adoptive À À À À À À cult to compare results obtained from different mouse models of transfer of cultured MCs from Apoe / mice into Apoe / Fcer1a /

AAA and in different experimental settings. In peri-aortic CaCl2 mice significantly enlarged AAA lesion sizes (Fig 4B) and increased injury-induced experimental AAAs, absence of CD4+ T cells or Th1 plasma IL6 levels (Fig 7D). The same adoptive transfer of MCs may cytokine IFN-c suppressed AAA formation. Intraperitoneal adminis- also change MC-specific proteases, although we are currently unable tration of IFN-c partially reversed AAA formations in CD4+ T-cell- to quantify mouse plasma chymase and tryptase due to a lack of suit- deficient mice (Xiong et al, 2004). In contrast, MHC mismatched able ELISA kits and to the complexity of different mouse chymase aortic transplantation-induced AAAs rely on Th2 cell functions. (mouse mast cell protease [mMCP]-1, -2, -4, -5, and -9) and tryptase Allografts from IFN-c receptor-deficient mice developed severe (mMCP-6, -7, and -11) isoforms. AAA, which can be blocked by anti-IL4 antibody or compound defi- IgE activation of macrophages leads to macrophage apoptosis and ciency of IFN-c receptor and IL4 (Shimizu et al, 2004). Therefore, release of IL6, MCP-1, and possibly other pro-inflammatory both Th1 and Th2 cells may contribute to AAA formation differ- mediators (Wang et al, 2011). To our surprise, adoptive transfer of À À ently, depending on the experimental model. In Ang-II infusion- macrophages from Apoe / mice did not significantly increase À À À À induced AAAs, however, complete elimination of all mature T cells plasma IL6 levels in Apoe / Fcer1a / mice (Fig 7E). Consistently, À À and B cells in Rag-1–deficient mice did not affect AAA incidence or outer aortic diameters did not increase significantly in Apoe / À À À À maximal aortic diameters in male or female mice, compared with Fcer1a / mice that received macrophages from Apoe / mice those with sufficient (Uchida et al,2010).Thisstudyby (Fig 5B). These observations reaffirm that IgE regulation of plasma Uchida et al may not definitively answer whether T cells or B cells IL6 correlated with AAA growth and also implicate IgE actions on participate in AAAs, but among T cells or B cells, some may promote macrophages, including IL6 production, in experimental AAA forma- AAA growth and some may inhibit AAA growth. For example, tion, but for lesser extent than IgE actions on CD4+ T cells or MCs. innate-like B1 cells protect mice from diet-induced atherosclerosis IgE stimulation of vascular SMCs and ECs may also affect AAA (Sun et al, 2010) and may also protect mice from AAA formation, pathogenesis. We detected negligible IgE in SMC-rich regions in although currently no direct test exists for each T-cell or B-cell human AAAs (Fig 8C). In contrast, IgE colocalized with SMC

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 963 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Figure 8. IgE and FceR1 expression in human AAAs lesions. A Immunostaining of human AAA lesion consecutive frozen sections to detect IgE and FceR1 expression in CD68-positive macrophages and TUNEL reactivity. Mouse IgG and rabbit IgG were used as immunostaining negative controls for mouse anti-IgE and rabbit anti-FceR1 antibodies. B Immunofluorescent double staining of CD4 and IgE or FceR1 to detect IgE and FceR1 expression in CD4+ T cells. C, D IgE and a-actin immunofluorescent double staining in smooth muscle cells (SMC)-rich and SMC-sparse areas. EFceR1 and a-actin immunofluorescent double staining to detect FceR1-positive SMCs. FCD31 and IgE or FceR1 immunofluorescent double staining of human AAA lesions. Data information: Scale: 200 lm; insert scale: 60 lm. Arrows indicate IgE-positive or FceR1-positive CD4+ T cells, SMCs or endothelial cells (ECs).

paucicellular portions of human AAA lesions (Fig 8D), suggesting a et al, 2011). IgE-induced apoptosis of vascular cells may contribute contribution of IgE to SMC apoptosis (Wang et al, 2011). In human to SMC depletion in human AAA lesions (Fig 8D), and intimal AAA lesions, IgE also colocalized with ECs (Fig 8F). Our prior in erosion. vitro experiments from cultured human SMCs and ECs suggested In the inflammatory cell reconstitution tests in experimental that IgE promotes EC apoptosis and cytokine production (Wang AAAs, we used CD4+ Tcells,CD8+ T cells, MCs, and macrophages

964 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

À À À À À À from both Apoe / and Apoe / Fcer1a / mice. We assumed that reagents from Pointe Scientific (Canton, MI). Experimental aneu- À À the only difference between the cells from Apoe / mice and those rysms were quantified using the methods of Daugherty et al as used À À À À from Apoe / Fcer1a / mice was the expression of FceR1. IgE only in our earlier studies (Daugherty et al, 2000; Schulte et al, 2010; À À À À À À activates cells from Apoe / mice, not those from Apoe / Fcer1a / Zhang et al, 2012). The suprarenal maximal aortic diameter of each mice (Wang et al, 2011), but absence of FceR1 may affect the expres- aneurysm was measured after the peri-adventitial tissue was care- sion of any other gene product, which may confound our data inter- fully dissected from the aortic wall. AAA incidence was defined by pretation from all adoptive transfer experiments (Figs 3, 4, 5 and increase of suprarenal maximal aortic diameter greater than 50% of Supplementary Fig S9). We did not test this possibility in this study. the mean value from same-age mice that received saline alone À À Instead, we treated Apoe / mice with anti-IgE antibody while these (0.97 Æ 0.02 mm, n = 10), according to previously reported meth- mice received Ang-II infusion. This antibody should interrupt IgE ods (Wang et al, 2006). The percentage of AAA incidence and actions on all cells that express FceR1. Two doses of anti-IgE anti- post-Ang-II infusion mortality rate per group were determined. À À body to Ang-II-infused Apoe / mice effectively ablated plasma IgE Determination of statistical significances between the groups used and inhibited AAA growth and associated lesion inflammation, the two-tailed Fisher’s exact test. All animal procedures conformed apoptosis, microvascularization, and cell proliferation to the same with the Guide for the Care and Use of Laboratory Animals À À À À extents as in Ang-II-infused Apoe / Fcer1a / mice. These observa- published by the U.S. National Institutes of Health and were À À tions seem confirmative of IgE functions in AAA from Apoe / approved by the Harvard Medical School Standing Committee on À À Fcer1a / mice, but have clinical significance. In a similar experi- Animals (protocol # 03759). ment presented in our earlier study, we used MC-deficient mice to prove a role of MCs in AAAs. In addition, we used over-the-counter Mouse aortic tissue immunohistochemical analysis MC stabilizers and established an effective prevention of AAA growth in aortic elastase perfusion-induced AAAs (Sun et al, Aorta segments for immunohistochemistry were cut at the maximal 2007b). Anti-IgE antibodies such as (also known as suprarenal outer aortic diameter and embedded vertically with rhuMab-E25 or XolairTM) and (also known as TNX-901) optimal cutting temperature (OCT) compound, and at least 10–15 are currently approved by the U.S. Food and Drug Administration serial frozen sections covering the maximal dilated aorta were (FDA) for clinical use in patients with allergic asthma, food aller- prepared for immunohistochemical analysis as described previously gies, or other allergic conditions (Milgrom et al, 1999; Leung et al, (Schulte et al, 2010; Zhang et al, 2012). For those with similarly 2003; Busse et al, 2011). Effective prevention of experimental AAAs enlarged AAA diameter throughout the thoracic-abdominal aortas, by anti-IgE antibody from this study suggests a potential clinical we selected the segment at approximately the same distance from application of omalizumab and Talizumab among AAA patients. the renal artery as those of others with maximal AAA expansions. In cases with AAA lesions at multiple locations, we selected the larg- est lesion as close as possible to the same distance from the renal Materials and Methods artery as those of others with maximal AAA expansions. Slides of each sample from identical levels were used for staining with each Mouse AAA production and lesion characterization antibody. Serial cryostat cross-sections (6 lm) were used for immu- nostaining for macrophages (Mac-2, BD Biosciences, 1:1,000), T À À We crossbred Fcer1a / mice (C57BL/6, N9, provided by Marie- cells (CD4, 1:90; BD Biosciences; and CD8, 1:300; Abcam, Helene Jouvin and Jean-Pierre Kinet, Harvard Medical School) with Cambridge, MA), dendritic cells (CD11c, 1:40, BD Biosciences), À À Apoe / mice (C57BL/6, N11, The Jackson Laboratory, Bar Harbor, MHC-II (1:250; BD Biosciences), MCP-1 (1:50, BD Biosciences), elas- À À À À À À ME) to generate Apoe / Fcer1a / mice and Apoe / control mice. tin (Verhoeff–van Gieson, Sigma-Aldrich), collagen (0.1% Sirius All mice used in this study were littermates in a syngeneic C57BL/6 Red F3BA; Polysciences Inc., Warrington, PA), a-actin (SMC, 1:750; À À À À background. To induce AAAs in Apoe / Fcer1a / mice (n = 15) Sigma-Aldrich), Ki67 (cell proliferation marker, 1:500; Vector Labo- À À and Apoe / control mice (n = 30), anesthetized (200 mg/kg ratories, Inc., Burlingame, CA), and CD31 (angiogenesis marker, ketamine, 10 mg/kg xylazine, intraperitoneal) 2-month-old male 1:1,500; BD Biosciences). Lesion apoptotic cells were determined mice were infused with 1000 ng/kg/min Ang-II (Sigma-Aldrich, St. with the in situ apoptosis detection kit, according to the manufac- Louis, MO) subcutaneously delivered by Alzet model 2004 osmotic turer’s instructions (Millipore, Billerica, MA). Elastin degradation minipumps (DURECT Corp., Cupertino, CA) for 28 days while mice and media SMC accumulation were graded according to the grading consumed a high-fat diet (C12108; Research Diets, Inc., New Bruns- keys described previously (Sun et al, 2007b). T cells, Ki67-positive wick, NJ). Post-operative analgesia (buprenophine, 0.05 mg/kg/ cells, apoptotic-positive cells, dendritic cells, and CD31-positive 12 h, intraperitoneal) was administered every 12 h for 48 h. Mouse microvessels were counted blindly. Unlike aortic cross sections from body weights were recorded before and after Ang-II infusion. Mouse fixed tissues that allowed counting of most individual immuno- diastolic and systolic blood pressures and heart rates were deter- positive signals (Rateri et al, 2011), we prepared aortic cross mined using the CODA non-invasive blood pressure system (Kent sections from unfixed abdominal aortas optimized for immunohisto- Scientific Co., Torrington, CT). Mice were sacrificed with carbon chemical study, but not for reliable enumeration of all individual dioxide narcosis, followed by cardiac puncture blood collection. immuno-positive cells. We therefore measured macrophage- Plasma IgE, IL6, IFN-c, IL10, and IgE levels were determined by positive, MHC-II-positive, MCP-1-positive, and collagen-positive ELISA according to the manufacturer’s protocol (BD Biosciences, areas using computer-assisted image analysis software (Image-Pro San Jose, CA). Plasma total cholesterol, triglyceride, and Plus; Media Cybernetics, Bethesda, MD). Ang-II-induced AAAs often high-density lipoprotein (HDL) levels were determined using show regions of disrupted media (Rateri et al, 2011), we calculated

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 965 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

AAA lesion areas from regions with both intact and fragmented negative control. Mice received a second dose of the same antibody media (Schulte et al, 2010; Zhang et al, 2012). or IgG1 isotype 14 days after surgery. Mice were harvested 28 days after initial Ang-II infusion. BMMC, macrophage, culture and adoptive transfer T-cell real-time polymerase chain reaction (RT-PCR), FACS, immunoblot analysis CD4+ and CD8+ T cells (2.5 × 106/ml) were À À À À To prepare BMMCs, bone-marrow cells from Apoe / and Apoe / cultured in a complete medium (RPMI 1640 medium and 10% fetal À À Fcer1a / mice were cultured for 5–6 weeks in the presence of bovine serum) in anti-CD3 (1 lg/ml) mAb (BD Pharmingen) pre- mouse recombinant IL3 (PeproTech, Inc., Rocky Hill, NJ) and stem coated culture dishes. After treatment with different stimuli, includ- cell factor (PeproTech), as reported previously (Sun et al, 2007a,b). ing INF-c (20 ng/ml), TNF-a (10 ng/ml), IL6 (20 ng/ml), IgE BMMC purity (nearly 100%) was confirmed with FACS analysis (50 lg/ml), and Ang-II (100 nM), total cellular RNA was extracted using FITC-conjugated CD117 mAb (eBioscience, Inc., San Diego, using Qiagen RNA isolation kit. Equal amounts of RNA were reverse CA) (Sun et al, 2007a). Mouse thioglycolate-stimulated (3%, Sigma- transcribed, and quantitative PCR was performed in a single-color À À Aldrich) peritoneal macrophages were harvested from Apoe / and RT-PCR detection system (Stratagene, La Jolla, CA). The mRNA À À À À Apoe / Fcer1a / mice, cultured in plasma-free medium for 2 h, levels of FceR1-a,FceR1-b, and FceR1-c chains were normalized to suspension cells removed, adhesive macrophages harvested, and those of b-actin. macrophage purity examined using FACS with anti-CD11b-FITC and For flow cytometry, cells were stained with the proper combina- anti-F4/80-APC mAbs (Caltag Laboratories, Burlingame, CA). tion of antibodies and analyzed on a flow cytometer FC500 (Beck- Peripheral CD4+ and CD8+ T cells from splenocytes of 8-week-old man Coulter, Brea, CA). The following antibodies were used for T- À À À À À À Apoe / and Apoe / Fcer1a / mice were first purified using a cell flow cytometry: anti-CD3-APC, anti-CD4-PE, anti-FceR1a-biotin, Nylon column according to the manufacturer’s instructions (Poly- anti-CD8-Alex700 mAbs (all from eBioscience, Inc.). For surface sciences, Inc., Warrington, PA), followed by antibody/complement- staining, cells were incubated with antibodies for 20 min at 4°C. For mediated depletion of B cells and CD8+ T cells (purified I-Ab and intracellar staining, cells were fixed and permeabilized according to CD8a antibodies, BD Pharmingen) or B cells and CD4+ T cells (puri- the manufacturer’s instructions (eBioscience, Inc.) before adding the fied I-Ab and CD4 antibodies, BD Pharmingen). Purified CD4+ and anti-FceR1a-biotin antibody. Isotype controls were used to correct CD8+ T cells were further purified by MACS-sorting using mouse T- compensation and to confirm antibody specificity. All samples were cell isolation kits according to the manufacturer’s instructions (Milt- analyzed using flow cytometry on a FACSCalibur (BD Biosciences). enyi Biotec Inc., Auburn, CA). Cell purity was confirmed by FACS For immunoblot analysis, an equal amount of protein from each analysis using anti-CD3-APC, anti-CD4-PE, and anti-CD8-Alex700 cell type preparation was separated by SDS–PAGE, blotted, and mAbs (all from eBioscience, Inc.). detected with mAbs against mouse FceR1a (1:1,000; eBioscience, À À À À To reconstitute cells in vivo, 8-week-old male Apoe / Fcer1a / Inc.) and b-actin (used for protein loading control, 1:1,000; Santa À À mice were given in vitro prepared BMMCs (n = 18 for Apoe / cells Cruz Biotechnology Inc., Santa Cruz, CA). À À À À and n = 12 for Apoe / Fcer1a / cells), peritoneal macrophages À À À À À À (n = 14 for Apoe / cells and n = 10 for Apoe / Fcer1a / cells), T-cell survival, proliferation, and cytokine production À À À À À À CD4+ (n = 20 for Apoe / cells and n = 15 for Apoe / Fcer1a / À À À À cells), and CD8+ (n = 25 for Apoe / cells and n = 10 for Apoe / To test whether IgE affects T-cell survival and proliferation, we À À Fcer1a / cells) T cells by tail vein injection (1 × 107 cells per seeded different numbers of CD4+ or CD8+ T cells (2.5 × 105 to mouse). Unlike other methods of bone-marrow-cell transplantation, 1 × 106 per well) on a 96-well flat-bottomed tissue culture plate À À À À irradiating the Apoe / Fcer1a / recipient mice before receiving containing both anti-CD3 (1 lg/ml) and anti-CD28 (1 lg/ml) anti- donor cell intravenous transfer is not necessary, and donor cells bodies. Cells were treated with murine IgE (50 lg/ml; Sigma- migrate to most organs and injured aortas. Recipient mice were Aldrich) for 3 days. MTT assay was then used to detect cell survival, introduced to the Ang-II infusion-induced experimental AAAs 1 day according to the manufacturer’s instructions (Millipore). Medium after tail vein injection. Donor cells in AAA lesions were confirmed absorbance was read at 570 nm. ELISA (BD Biosciences) was then with immunostaining aortic sections with hamster anti-mouse used to determine IL12 levels from T-cell culture media. FceR1a mAb (1:100, eBioscience, Inc.) at harvesting. To minimize To detect cytokine production in CD4+ and CD8+ T cells, T cells inter-experimental variations, we used the Ang-II and minipump (2.5 × 106/ml) were cultured in a complete medium (RPMI 1640 from the same lot number and had the same investigator perform medium and 10% fetal bovine serum) in anti-CD3 mAb (1 lg/ml, implantation. We also planned the procedures on various experi- BD Pharmingen) pre-coated plates in the presence or absence of À À À À À À mental groups including Apoe / and Apoe / Fcer1a / control different doses of IgE (0–100 lg/ml; Sigma-Aldrich) for 2 days. Cells mice so that some mice in each group were done each day, rather were collected for total RNA preparation using the Qiagen RNA than having only one condition processed per day. isolation kit for RT-PCR (Bio-Rad, Hercules, CA) to determine mRNA levels of IL6, IFN-c, and IL10. Cell culture media were Anti-IgE antibody treatment collected to measure secreted cytokines by ELISA, according to the manufacturer’s instructions (BD Biosciences). À À Two-month-old male Apoe / mice received tail-vein injection of rat anti-mouse IgE mAb (n = 11) in a dose previously validated in Patient population and IgE ELISA mice (Coyle et al, 1996; Haile et al, 1999) (330 lg in 200 ll of saline per 25 g body weight, BD Pharmingen) 1 day before the surgery. In an ongoing randomized population-based screening trial for Matched rat IgG1 isotype (n = 10, BD Pharmingen) was used as AAAs, PAD, and hypertension in more than 50,000 men 65–

966 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

74 years of age in the mid-region of Denmark (Grøndal et al, 2010), The paper explained baseline plasma samples were taken consecutively at diagnosis of 476 AAA patients and 200 age-matched controls without AAA or Problem PAD. AAA was defined as having maximal aortic diameter greater We previously showed that immunoglobulin E (IgE) activates cultured than 30 mm, and PAD was defined as an ankle-brachial index (ABI) MCs and macrophages, both are important pro-inflammatory cells in lower than 0.90 or >1.4. Ankle systolic blood pressure was human and experimental AAAs. However, it remains unknown measured as previously validated and reported (Joensen et al, whether IgE activation of these cells or other inflammatory cells – contributes to AAA pathogenesis, and whether it is possible that inter- 2008), and maximal anterior posterior diameter of the infrarenal ruption of this IgE activity ameliorates AAA development. aorta was measured in the peak of the systole from inner edge to inner edge of the aorta. Patients with AAAs less than 50 mm were Results offered annual control scans by the screening team; patients with This study demonstrates that not only MCs and macrophages, but AAAs measuring 50 mm or larger were referred for a computed also CD4+ and CD8+ T cells express IgE receptor FceR1. IgE induces 4+ 6 c tomography (CT) scan and vascular surgical evaluation. The inter- CD T-cell production of pro-inflammatory cytokines IL and IFN- , but reduces their production of anti-inflammatory cytokine IL10. observer variation of aortic diameter measurements was 1.52 mm Genetic depletion of this IgE receptor protects mice from chronic et al (Grøndal , 2012). Growth rates of small AAAs in patients kept angiotensin II infusion-induced AAAs. We are also able to reverse under surveillance were calculated by individual linear regression reduced AAAs and plasma IL6 levels in FceR1-deficient mice by repopu- analysis, using all observations. Blood samples were centrifuged at lating CD4+ T cells, MCs, and macrophages from FceR1-sufficient mice e 1 3,000 g for 12 min, aliquoted, and stored at À80°C until analysis but not those from Fc R -deficient mice. Relevant to clinical implica- tion, we demonstrate that biweekly intravenous administration of an was performed. All subjects gave informed consent before partici- anti-IgE monoclonal antibody reduces plasma IgE by more than 70% pating, and the Local Ethics Committee of the Viborg Hospital, and yields similar suppression of AAAs to those in FceR1-deficient mice. Denmark, approved the study, which was performed in accordance In humans, we also discover that AAA patients have higher plasma IgE with the Helsinki Declaration. The Partners Human Research levels than those without AAAs. Committee (Boston, MA) also approved the use of non-coded human samples. Plasma IgE levels were determined using the Impact Human IgE Ready-SET-Go! according to the manufacturer’s instruc- This study establishes an important role of IgE in AAA formation by activating inflammatory cells, including CD4+ T cells, MCs, macro- tions (eBioscience, Inc.). phages, and possibly other untested cell types. Reduced AAAs in mice that are genetic deficient for IgE receptor FceR1 or that receive treat- Human AAA lesion immunohistology ment with an anti-IgE monoclonal antibody suggest a future therapeu- tic potential of this aortic disease using anti-IgE antibodies, such as Immunohistochemistry with mouse anti-human IgE (1:70, Southern omalizumab and Talizumab that are currently used widely among patients with asthma or other allergic diseases. Biotech, Birmingham, AL) and rabbit anti-human FceR1 (1:300; Santa Cruz Biotechnology) and their colocalization with apoptosis (TUNEL, Apoptosis Tag kit; Millipore) was performed on acetone fixed frozen sections from human AAA (n = 11) and control normal curve was 0.5. If the lowest 95% confidence limit for the area under aortas (n = 5) by avidin/biotin method. Specimens obtained by the curve was above 0.5, a significant predictive test was present. protocols approved by the Human Investigation Review Committee The optimal cut-off points were determined, and the respective sensi- at the Brigham and Women’s Hospital were immediately immersed tivity and specificity were calculated. The associations of plasma IgE in saline, embedded in OCT, and stored at À80°C freezer. To local- were correlated to maximal aortic diameter, lowest ABI, and AAA ize IgE and FceR1 to vascular SMCs, ECs, and CD4+ T cells, we growth rate by Spearman’s correlation test. Because of relatively performed double immunofluorescent staining (n = 11) with mouse small sample sizes and often skewed data distribution, we selected anti-IgE (1:10) or rabbit anti-FceR1 (1:80) antibodies incubated over- the non-parametric Mann–Whitney U test for paired data sets and night, followed by anti-mouse or anti-rabbit Alexa-555 (Invitrogen, one-way ANOVA with post-hoc Bonferroni test was used for compar- Grand Island, NY). Sections were washed, blocked with 4% of ison among three or more groups to examine statistical significance appropriate normal serum, and incubated with mouse anti-a-actin for all data from cultured cells and experimental AAAs. Fisher’s exact (1:40, HHF35; Enzo-Diagnostics, Syosset, NY), CD31 (1:40, Dako, test was used to compare the differences in AAA incidence and post- Carpinteria, CA), and CD4 (1:20, Dako), followed by anti-mouse Ang-II mortality. P < 0.05 was considered statistically significant. Alexa-488 (Invitrogen). Supplementary information for this article is available online: Statistical analysis http://embomolmed.embopress.org

One sample Kolmogorov–Smirnov test and probability plot (not Acknowledgements shown) were used to determine whether human plasma IgE levels The authors thank Mrs. Eugenia Shvartz for technical assistance, Ms. Sara were normally distributed, and compared between controls and Karwacki for editorial assistance, and Drs. Marie-Helene Jouvin and Jean-Pierre cases with Student’s t-test and the Mann–Whitney U test. ROC curve Kinet of Beth Israel Deaconess Medical Center and Harvard Medical School for analyses were performed non-parametrically to test the predictive providing the FceR1a-deficient mice. This study is supported by grants from value of the test, concerning the prediction of AAA cases. For analy- the National Institutes of Health (HL60942,HL81090,HL88547 to GPS; ses of the ROC curves, the null hypothesis was that the test HL34636,HL80472 to PL), and by an American Heart Association Established performed similarly to the diagonal line—i.e. the area under the Investigator Award (0840118N, to GPS).

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 967 EMBO Molecular Medicine IgE actions in abdominal aortic aneurysms Jing Wang et al

Author contributions Grøndal N, Bramsen MB, Thomsen MD, Rasmussen CB, Lindholt JS (2012) The JW, GKS, MAS, HC, AH, and YW performed the mouse experiments and human cardiac cycle is a major contributor to variability in size measurements of sample analysis. JSL, MX, and JAW collected human specimens and data analy- abdominal aortic aneurysms by ultrasound. Eur J Vasc Endovasc Surg 43: sis. MX, and NX helped with FACS analysis. PL helped with manuscript writing 30 – 33 and data explanation. GPS designed the experiments and wrote the manu- Haile S, Lefort J, Eum SY, Dumarey C, Huerre M, Heusser C, Vargaftig BB script. (1999) Suppression of immediate and late responses to antigen by a non-anaphylactogenic anti-IgE antibody in a murine model of asthma. Conflict of interest Eur Respir J 13: 961 – 969 The authors declare that they have no conflict of interest. Heikkilä HM, Lätti S, Leskinen MJ, Hakala JK, Kovanen PT, Lindstedt KA (2008) Activated mast cells induce endothelial cell apoptosis by a combined action of chymase and tumor necrosis factor-alpha. Arterioscler Thromb References Vasc Biol 28: 309 – 314 Huang M, Pang X, Letourneau R, Boucher W, Theoharides TC (2002) Acute Andoh T, Kuraishi Y (2004) Expression of Fc epsilon receptor I on primary stress induces cardiac mast cell activation and histamine release, effects sensory neurons in mice. NeuroReport 15: 2029 – 2031 that are increased in Apolipoprotein E knockout mice. Cardiovasc Res 55: Baniyash M, Eshhar Z, Rivnay B (1986) Relationships between on IgE 150 – 160 recognized by defined monoclonal antibodies and by the FC epsilon Ishizaka T, Ishizaka K, Conrad DH, Froese A (1978) A new concept of receptor on basophils. J Immunol 136: 588 – 593 triggering mechanisms of IgE-mediated histamine release. J Allergy Clin Boesiger J, Tsai M, Maurer M, Yamaguchi M, Brown LF, Claffey KP, Dvorak HF, Immunol 61: 320 – 330 Galli SJ (1998) Mast cells can secrete vascular permeability factor/ vascular Joensen JB, Juul S, Abrahamsen J, Henneberg EW, Lindholt JS (2008) Doppler endothelial cell growth factor and exhibit enhanced release after ultrasound compared with strain gauge for measurement of systolic ankle immunoglobulin E-dependent upregulation of fc epsilon receptor I blood pressure. Angiology 59: 296 – 300 expression. J Exp Med 188: 1135 – 1145 Katoh N, Kraft S, Wessendorf JH, Bieber T (2000) The high-affinity IgE Bot I, Biessen EA (2011) Mast cells in atherosclerosis. Thromb Haemost 106: receptor (FcepsilonRI) blocks apoptosis in normal human monocytes. J Clin 820 – 826 Invest 105: 183 – 190 Busse WW, Morgan WJ, Gergen PJ, Mitchell HE, Gern JE, Liu AH, Gruchalla RS, Kawakami T, Galli SJ (2002) Regulation of mast-cell and basophil function Kattan M, Teach SJ, Pongracic JA et al (2011) Randomized trial of and survival by IgE. Nat Rev Immunol 2: 773 – 786 omalizumab (anti-IgE) for asthma in inner-city children. N Engl J Med 364: King VL, Lin AY, Kristo F, Anderson TJ, Ahluwalia N, Hardy GJ, Owens AP 3rd, 1005 – 1015 Howatt DA, Shen D, Tager AM et al (2009) -gamma and the Coyle AJ, Wagner K, Bertrand C, Tsuyuki S, Bews J, Heusser C (1996) Central interferon-inducible chemokine CXCL10 protect against aneurysm role of immunoglobulin (Ig) E in the induction of lung eosinophil formation and rupture. Circulation 119: 426 – 435 infiltration and T helper 2 cell cytokine production: inhibition by a Kokkinopoulos D, Perez S, Sotiriadou R, Stinios J, Papamichail M (1992) The non-anaphylactogenic anti-IgE antibody. J Exp Med 183: 1303 – 1310 use of nylon wool for the isolation of T subpopulations. J Daugherty A, Manning MW, Cassis LA (2000) Angiotensin II promotes Immunol Methods 154: 1 – 6 atherosclerotic lesions and aneurysms in apolipoprotein E-deficient mice. J Kovanen PT (2007) Mast cells and degradation of pericellular and extracellular Clin Invest 105: 1605 – 1612 matrices: potential contributions to erosion, rupture and intraplaque den Dekker WK, Tempel D, Bot I, Biessen EA, Joosten LA, Netea MG, van der haemorrhage of atherosclerotic plaques. Biochem Soc Trans 35: 857 – 861 Meer JW, Cheng C, Duckers HJ (2012) Mast cells induce vascular smooth Lappalainen H, Laine P, Pentikäinen MO, Sajantila A, Kovanen PT (2004) Mast muscle cell apoptosis via a toll-like receptor 4 activation pathway. cells in neovascularized human coronary plaques store and secrete basic Arterioscler Thromb Vasc Biol 32: 1960 – 1969 fibroblast growth factor, a potent angiogenic mediator. Arterioscler Thromb Dileepan KN, Stechschulte DJ (2006) Endothelial cell activation by mast cell Vasc Biol 24: 1880 – 1885 mediators. Methods Mol Biol 315: 275 – 294 Leung DY, Sampson HA, Yunginger JW, Burks AW Jr, Schneider LC, Wortel CH, Dombrowicz D, Quatannens B, Papin JP, Capron A, Capron M (2000) Davis FM, Hyun JD, Shanahan WR Jr; Avon Longitudinal Study of Parents Expression of a functional Fc epsilon RI on rat eosinophils and and Children Study Team (2003) Effect of anti-IgE therapy in patients macrophages. J Immunol 165: 1266 – 1271 with . N Engl J Med 348: 986 – 993. Dvorak AM, Schulman ES, Peters SP, MacGlashan DW Jr, Newball HH, Leuschner F, Rauch PJ, Ueno T, Gorbatov R, Marinelli B, Lee WW, Dutta P, Schleimer RP, Lichtenstein LM (1985) Immunoglobulin E-mediated Wei Y, Robbins C, Iwamoto Y et al (2012) Rapid monocyte kinetics in degranulation of isolated human lung mast cells. Lab Invest 53: acute myocardial infarction are sustained by extramedullary 45 – 56 monocytopoiesis. J Exp Med 209: 123 – 137 Gilfillan AM, Tkaczyk C (2006) Integrated signalling pathways for mast-cell Mancardi DA, Iannascoli B, Hoos S, England P, Daëron M, Bruhns P (2008) activation. Nat Rev Immunol 6: 218 – 230 FcgammaRIV is a mouse IgE receptor that resembles macrophage Grayson MH, Cheung D, Rohlfing MM, Kitchens R, Spiegel DE, Tucker J, FcepsilonRI in humans and promotes IgE-induced lung inflammation. J Battaile JT, Alevy Y, Yan L, Agapov E et al (2007) Induction of high-affinity Clin Invest 118: 3738 – 3750 IgE receptor on lung dendritic cells during viral infection leads to mucous Milgrom H, Fick RB Jr, Su JQ, Reimann JD, Bush RK, Watrous ML, Metzger WJ cell metaplasia. J Exp Med 204: 2759 – 2769 (1999) Treatment of allergic asthma with monoclonal anti-IgE antibody. Grøndal N, Søgaard R, Henneberg EW, Lindholt JS (2010) The Viborg Vascular rhuMAb-E25 Study Group. N Engl J Med 341: 1966 – 1973 (VIVA) screening trial of 65–74 year old men in the central region of Milner JD, Ward JM, Keane-Myers A, Paul WE (2007) Lymphopenic mice Denmark: study protocol. Trials 11: 67 reconstituted with limited repertoire T cells develop severe, multiorgan,

968 EMBO Molecular Medicine Vol 6 |No7 | 2014 ª 2014 The Authors Jing Wang et al IgE actions in abdominal aortic aneurysms EMBO Molecular Medicine

Th2-associated inflammatory disease. Proc Natl Acad Sci U S A 104: Sun J, Sukhova GK, Zhang J, Chen H, Sjöberg S, Libby P, Xia M, Xiong N, Gelb 576 – 581 BD, Shi GP (2012) Cathepsin K deficiency reduces elastase Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M, Otsu M, perfusion-induced abdominal aortic aneurysms in mice. Arterioscler Hara K, Ueki K, Sugiura S et al (2009)CD8+ effector T cells contribute to Thromb Vasc Biol 32: 15 – 23 macrophage recruitment and adipose tissue inflammation in obesity. Nat Swedenborg J, Mäyränpää MI, Kovanen PT (2011) Mast cells: important Med 15: 914 – 920 players in the orchestrated pathogenesis of abdominal aortic aneurysms. Qin Y, Cao X, Guo J, Zhang Y, Pan L, Zhang H, Li H, Tang C, Du J, Shi GP Arterioscler Thromb Vasc Biol 31: 734 – 740 (2012) Deficiency of attenuates angiotensin II-induced Theoharides TC, Kempuraj D, Tagen M, Conti P, Kalogeromitros D (2007) abdominal aortic aneurysm formation in apolipoprotein E-deficient mice. Differential release of mast cell mediators and the pathogenesis of Cardiovasc Res 96: 401 – 410 inflammation. Immunol Rev 217: 65 – 78 Rateri DL, Howatt DA, Moorleghen JJ, Charnigo R, Cassis LA, Daugherty A Thompson RW, Curci JA, Ennis TL, Mao D, Pagano MB, Pham CT (2006) (2011) Prolonged infusion of angiotensin II in apoE(-/-) mice promotes Pathophysiology of abdominal aortic aneurysms: insights from the macrophage recruitment with continued expansion of abdominal aortic elastase-induced model in mice with different genetic backgrounds. Ann N aneurysm. Am J Pathol 179: 1542 – 1548 Y Acad Sci 1085: 59 – 73 Razin E, Mencia-Huerta JM, Stevens RL, Lewis RA, Liu FT, Corey E, Austen KF Uchida HA, Kristo F, Rateri DL, Lu H, Charnigo R, Cassis LA, Daugherty A (1983) IgE-mediated release of C4, chondroitin sulfate E (2010) Total lymphocyte deficiency attenuates AngII-induced proteoglycan, beta-hexosaminidase, and histamine from cultured bone atherosclerosis in males but not abdominal aortic aneurysms in apoE marrow-derived mouse mast cells. J Exp Med 157: 189 – 201 deficient mice. Atherosclerosis 211: 399 – 403 Rohde LE, Arroyo LH, Rifai N, Creager MA, Libby P, Ridker PM, Lee RT (1999) Vucevic D, Maravic-Stojkovic V, Vasilijic S, Borovic-Labudovic M, Majstorovic I, Plasma concentrations of -6 and abdominal aortic diameter Radak D, Jevtic M, Milosavljevic P, Colic M (2012) Inverse production of among subjects without aortic dilatation. Arterioscler Thromb Vasc Biol 19: IL-6 and IL-10 by abdominal aortic aneurysm explant tissues in culture. 1695 – 1699 Cardiovasc Pathol 21: 482 – 489 Schulte S, Sun J, Libby P, MacFarlane L, Sun C, Lopez-Ilasaca M, Shi GP, Wang YX, Cassis LA, Daugherty A (2006) Angiotensin II-induced aortic Sukhova GK (2010) Cystatin C deficiency promotes inflammation in aneurysms. In A Handbook of Mouse Models of Cardiovascular Disease,Xu angiotensin II-induced abdominal aortic aneurisms in atherosclerotic Q (ed.), pp 125 – 135. Chichester, England: John Wiley & Sons. mice. Am J Pathol 177: 456 – 463 Wang J, Cheng X, Xiang MX, Alanne-Kinnunen M, Wang JA, Chen H, He A, Sun Shibaki A, Ohkawara A, Shimada S, Ra C, Aiba S, Cooper KD (1996) Expression, X, Lin Y, Tang TT et al (2011) IgE stimulates human and mouse arterial but lack of calcium mobilization by high-affinity IgE Fc epsilon receptor I cell apoptosis and cytokine expression and promotes atherogenesis in on human epidermal and dermal Langerhans cells. Exp Dermatol 5: Apoe-/- mice. J Clin Invest 121: 3564 – 3577 272 – 278 Xiong W, Zhao Y, Prall A, Greiner TC, Baxter BT (2004) Key roles of CD4+ T Shimizu K, Shichiri M, Libby P, Lee RT, Mitchell RN (2004)Th2-predominant cells and IFN-gamma in the development of abdominal aortic aneurysms inflammation and blockade of IFN-gamma signaling induce aneurysms in in a murine model. J Immunol 172: 2607 – 2612 allografted aortas. J Clin Invest 114: 300 – 308 Xiong W, Knispel R, MacTaggart J, Greiner TC, Weiss SJ, Baxter BT (2009) Sun J, Sukhova GK, Wolters PJ, Yang M, Kitamoto S, Libby P, MacFarlane LA, Membrane-type 1 matrix metalloproteinase regulates Mallen-St Clair J, Shi GP (2007a) Mast cells promote atherosclerosis by macrophage-dependent elastolytic activity and aneurysm formation in releasing proinflammatory cytokines. Nat Med 13: 719 – 724 vivo. J Biol Chem 284: 1765 – 1771 Sun J, Sukhova GK, Yang M, Wolters PJ, MacFarlane LA, Libby P, Sun C, Zhang Xu JM, Shi GP (2012) Emerging role of mast cells and macrophages in Y, Liu J, Ennis TL et al (2007b) Mast cells modulate the pathogenesis of cardiovascular and metabolic diseases. Endocr Rev 33: 71 – 108 elastase-induced abdominal aortic aneurysms in mice. J Clin Invest 117: Zhang J, Sun J, Lindholt JS, Sukhova GK, Sinnamon M, Stevens RL, Adachi R, Libby 3359 – 3368 P, Thompson RW, Shi GP (2011) Mast cell tryptase deficiency attenuates Sun J, Zhang J, Lindholt JS, Sukhova GK, Liu J, He A, Abrink M, Pejler G, mouse abdominal aortic aneurysm formation. Circ Res 108: 1316 – 1327 Stevens RL, Thompson RW et al (2009) Critical role of mast cell chymase Zhang J, Chen H, Liu L, Sun J, Shi MA, Sukhova GK, Shi GP (2012) Chemokine in mouse abdominal aortic aneurysm formation. Circulation 120: 973 – 982 (C-C motif) receptor 2 mediates mast cell migration to abdominal aortic Sun J, Hartvigsen K, Chou MY, Zhang Y, Sukhova GK, Zhang J, Lopez-Ilasaca M, Diehl CJ, aneurysm lesions in mice. Cardiovasc Res 96: 543 – 551 Yakov N, Harats D et al (2010) Deficiency of antigen-presenting cell invariant chain reduces atherosclerosis in mice. Circulation 122: 808 – 820 License: This is an open access article under the Sun J, Sukhova GK, Zhang J, Chen H, Sjöberg S, Libby P, Xiang M, Wang J, terms of the Creative Commons Attribution 4.0 Peters C, Reinheckel T et al (2011) Cathepsin L activity is essential to License, which permits use, distribution and reproduc- elastase perfusion-induced abdominal aortic aneurysms in mice. tion in any medium, provided the original work is Arterioscler Thromb Vasc Biol 31: 2500 – 2508 properly cited.

ª 2014 The Authors EMBO Molecular Medicine Vol 6 |No7 | 2014 969